论文信息 - A Critical Study of Efficient Multi-core EM Clustering

A Critical Study of Efficient Multi-core EM Clustering

state-of-the-art computer hardware is coming with multi- core processors. Even mobile phones are coming with dual- core processors. OpenMP is one technology supporting parallel programming on multi-core shared memory systems with the help of threads. In this paper, we observed the execution times Serial EM Clustering running on single-core and Parallel EM Clustering methods using OpenMP on I3 system. Observations are made varying number of threads, samples, dimensions and clusters. The results show that OpenMP Lower Triangular Canonical Form with Forward Substitution and Winograd's approach (OLFW) EM gives a considerable speed-up of 2.7 over serial standard EM.

N. B. Venkateswarlu | J. V. R. Murthy | D. J. Nagendra Kumar

[1] Dinesh B. Kulkarni,et al. Convex Hull Using K-Means Clustering in Hybrid (MPI/OpenMP) Environment , 2010, 2010 International Conference on Computational Intelligence and Communication Networks.

[2] M.D. Jones,et al. Hybrid MPI-OpenMP Programming for Parallel OSEM PET Reconstruction , 2006, IEEE Transactions on Nuclear Science.

[3] E. V. Prasad,et al. A Critical Performance Study of Memory Mapping on Multi- Core Processors: An Experiment with k-means Algorithm with Large Data Mining Data Sets , 2010 .

[4] Ian Witten,et al. Data Mining , 2000 .

[5] D. J. N. Kumar,et al. Some Studies of Expectation Maximization Clustering Algorithm to Enhance Performance , 2011 .

[6] M.D. Jones,et al. Parallel programming for OSEM reconstruction with MPI, OpenMP, and hybrid MPI-OpenMP , 2004, IEEE Symposium Conference Record Nuclear Science 2004..

[7] N. B. Venkateswarlu,et al. Fast Expectation Maximization Clustering Algorithm , 2012 .

[8] D. Rubin,et al. Maximum likelihood from incomplete data via the EM - algorithm plus discussions on the paper , 1977 .